1

I'm attempting to recover a 7 Drive RAID6 array from a failed Thecus NAS. I've been able to get the drives and access to the data through an Ubuntu machine I setup, the problem being that the transfer rates off of the raid are painfully slow (~500Kb/s - 1.2Mbs/).

I've discovered that one of the drives appears to be degraded, and am guessing that is probably the root of the problem. When performing a "mdadm --detail /dev/md0" I get the following results:

/dev/md0:
        Version : 1.2
  Creation Time : Tue May  7 15:39:33 2013
     Raid Level : raid6
     Array Size : 14638110720 (13959.99 GiB 14989.43 GB)
  Used Dev Size : 2927622144 (2792.00 GiB 2997.89 GB)
   Raid Devices : 7
  Total Devices : 6
    Persistence : Superblock is persistent

    Update Time : Thu Feb  8 08:02:27 2018
          State : clean, degraded 
 Active Devices : 6
Working Devices : 6
 Failed Devices : 0
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 64K

           Name : N7700PRO:0
           UUID : 7169575c:a8d508eb:dea20994:ee2351ef
         Events : 64278

    Number   Major   Minor   RaidDevice State
       7       8      130        0      active sync   /dev/sdi2
       2       0        0        2      removed
       2       8       82        2      active sync   /dev/sdf2
       3       8       34        3      active sync   /dev/sdc2
       4       8       50        4      active sync   /dev/sdd2
       5       8        2        5      active sync   /dev/sda2
       6       8       18        6      active sync   /dev/sdb2

I've got a spare drive on-hand for a failed drive, but I'm not entirely sure how to add it into the array and repair it. I've pulled the bad drive out of the system, and plugged the spare in it's place, but when performing the mdadm --detail I'm getting the same results as with the original drive in place.

I believe the command to add a drive is just

mdadm --add /dev/md0 <new_disk>

But, I'm unsure how to get the path for the new disk since it isn't appearing in the list, I wasn't seeing any information in the disk utility that matches the "/dev/sdx2" format either to lend any clues for the command.

I've also got all of the SATA ports on the Motherboard occupied at this point, and am wondering if that could be part of the issue as well? I'm not really sure, but here are the details of the machine as it sits-

  • 7x3TB WD REDS (RAID Drives)
  • 1x2TB WD Green (OS)
  • Asus Sabertooth 990fx r2
  • 16GB DDR3
  • AMD FX 8350
  • AMD 7870
  • XFX 850w PSU

Output from ls /dev/sd?; some investigation says that it looks like the new drive is /dev/sdg

/dev/sda  /dev/sdc  /dev/sde  /dev/sdg  /dev/sdi
/dev/sdb  /dev/sdd  /dev/sdf  /dev/sdh

Output from mount | awk '$3=="/"'

/dev/sdh1 on / type ext4 (rw,relatime,errors=remount-ro,data=ordered)

Let me know if there is any further information you may need, appreciate any and all assistance in this.

1

The disk layout is /dev/sdXN where X is a letter in the range [a-z] and N is a number in the range [1-9]. Each disk is represented as /dev/sdX so that's what you need to use to find the new disk. The N is the partition (slice) number; your RAID is expecting to use partition 2 on each disk so you need to find out what the disk layout is, and replicate that onto the new disk. Finally you can then add the partition to your RAID and let it rebuild.

  1. Identify the new disk

    You have said that it's /dev/sdg.

  2. Replicate the disk partition table

    It must be GPT because you are using 3TB disks (MBR works only for disks up to 2TB). We will replicate the partition table from /dev/sda onto the new disk /dev/sdg, remembering to generate new UUIDs along the way:

    sgdisk --replicate=/dev/sdg /dev/sda
    sgdisk --randomize-guids /dev/sdg
    

    If you don't have sgdisk installed you can find it in the gdisk package (Debian, Ubuntu, CentOS, etc.).

  3. Add the newly partitioned disk into the RAID array

    mdadm --add /dev/md0 /dev/sdg2
    

    Don't forget to let it rebuild (see cat /proc/mdstat for status details)

I would strongly recommend that you read the man page for sgdisk and mdadm to ensure that the commands I have suggested will indeed do what I have described and you would expect. If you lose a second disk from your RAID6 array you won't have any redundancy left.

  • You are a Rock star my friend. Thank you so much. – zroberts Feb 8 '18 at 17:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.